CONDITIONAL EUROBONDS AND

EUROZONE REFORM John Muellbauer, INET at Oxford

OENB workshop ‘Towards a genuine economic and monetary union’,

Vienna, 10-11 September, 2015

OBJECTIVES

• Reduce the Euro-area policy focus on fiscal austerity.

• Switch to a focus on improving competitiveness and productivity (hence growth).

• Simultaneously promote fiscal decentralisation and sensible fiscal policy with national incentives.

• Achieve all this through rules-based risk spreads for individual countries’ Eurobonds.

• Early version in 2011 http://www.voxeu.org/epubs/cepr-reports/resolving-eurozone-crisis-time-conditional-eurobonds

2

INTERNAL IMBALANCES

• Despite spotty improvements, internal imbalances in the common currency area remain a major problem:

o divergent fiscal discipline;

o profound differences in labour market, credit, and housing institutions;

o past failures in financial regulation have left a burden;

• This is visible in divergences in:

o competitiveness, government, and private debt to GDP ratios, balance of payments deficits, housing market experiences.

3

RESOLUTION?

• Fundamental issue to resolve:

o Eurozone lacks the democratic institutions for a fiscal union. Fiscal union without draconian centralisation may pose dangers of perverse incentives.

• Create the right incentives:

o mixture of carrots/sticks to enable the poorly performing economies to return to economic growth and avoid a future existential crisis;

o discourage moral hazard;

o arrive at a fair distribution of burden sharing between taxpayers in different countries and holders of sovereign and bank bonds.

4

CONDITIONAL EUROBONDS

• Conditional eurobonds:

o Eurobonds with a collective underwriting guarantee which eliminates the country risk faced by investors;

o Administratively set spreads determine the annual side-payments the riskier countries (Ireland, Italy, Portugal, Spain) pay into a common fund.

o Spreads would compensate the taxpayers in less risky countries for their risk in underwriting the bonds of the riskier countries.

5

RESOLUTION?

• Conditional eurobonds with risk premia based on economic fundamentals could:

• Create the incentives for the fundamental structural reforms still outstanding.

• Limit the country sovereign debt risk faced by investors (other than in the special case of Greece).

• Avoid the ‘transfer union’ feared by Germany.

6

CONDITIONAL EUROBONDS

• Conditional eurobonds:

o All new borrowing would be in the form of eurobonds. Only countries not under a troika –supervised programme permitted to issue eurobonds.

o The spreads would be set conditional on a set of clear performance targets ratified by a new European monetary and fiscal authority (EMFA).

o As economic fundamentals improve, fiscal authorities know that borrowing costs will decline.

7

CONFUSING MARKET SIGNALS

• Current spreads of 10 year bond yields relative to Germany have been depressed below economic fundamentals by QE programme.

• Market signals also worked badly in previous years:

• Before 2009, markets paid far too little attention to underlying fundamentals.

• Between 2009 and 2013, markets were distorted by panics, fear of euro-area break up and fluctuations in risk appetite.

• Policy makers don’t know how to interpret the signals.

8

RISK PREMIA BASED ON FUNDAMENTALS

o Create strong incentive for the right reform priorities .

o This makes conditional eurobonds quite different from conventional eurobonds.

o Kopf (2011) and Gros (2011): conventional eurobonds suffer from incentive problems, creating the risk of a future, even larger crisis.

9

ECONOMETRIC MODEL TO EXTRACT FUNDAMENTALS

o Model for 10-year yield spreads relative to Germany needs to take account of:

o inattention before 2007

o Growing attention to fundamentals from 2007 to 2010

o Temporary amplification in market panics and falls in risk appetite

o Effect of QE from late 2014

o Model in

http://www.inet.ox.ac.uk/files/publications/ukmfeat4.pdf

updates OXREP https://ideas.repec.org/a/oup/oxford/v29y2013i3p610-645.html

10

A NEW ECONOMETRIC MODEL

• DATA

o Economic fundamentals data: quarterly frequencies.

o Panel model with quarterly data on spreads in 10-year sovereign bond yields relative to Germany for 9 other euro area countries.

o Omit Greece: market did not have a realistic assessment of its fundamental data until 2010.

o Include Ireland and Portugal: up to the quarters preceding their bailouts (in 2010Q2 and 2010Q3).

o Terms of the rescue packages supervised by the ‘troika’ affected spreads. Ireland then included again for 2014Q1 observation.

11

A NEW ECONOMETRIC MODEL

• STRUCTURAL CHANGE

o PHASE 1: beginning in 2007Q3:

o Linked with the drying up of money markets, initially triggered by losses in money market funds, partly invested in sub-prime linked securities.

o PHASE 2: beginning in the second half of 2009 with full ‘attentiveness’ being reached by the end of 2010.

o Triggered by increasing worries about sovereign debt and imbalances(dramatised by Greece).

12

A NEW ECONOMETRIC MODEL

• STRUCTURAL CHANGE

o The two phases are handled by a linear combination of two dummy variables making a smooth transition from zero to one between mid-2007 and the beginning of 2009 and from then to the last quarter of 2010.

o Rising attentiveness:

13

A NEW ECONOMETRIC MODEL

• OVERREACTION OR AMPLIFICATION

o First: by introducing sensitivity to global risk appetite.

o Second: by use of dummy variables to capture euro -area specific ‘alarm’.

• The alarm function is influenced both by stated policy of the ECB and other euro area policy makers and by ECB interventions in the sovereign bond markets.

• The model thus includes a time-varying scale factor consisting of ‘attention’ plus ‘alarm’ which amplifies the scalar function of the basic economic drivers of each country’s long-run spreads relative to Germany.

14

A NEW ECONOMETRIC MODEL

o ‘Longterm’ solution for the yield spreads can therefore be defined by

yit = f(xit)(attentiont + alarmt) (2)

o Attention=1 from 2010Q4; steady state value of alarm=0.

o In a post-2010 steady state, we have

yit = f(xit)

o Slightly augmented partial adjustment model for the quarterly change in spreads captures the dynamics well.

o Speed of adjustment: 25% per quarter; so that around 70% of the adjustment is complete after a year.

15

A NEW ECONOMETRIC MODEL

• OVERREACTION OR AMPLIFICATION

o Estimated value of (attentiont + alarmt) shown; includes plot evaluated when index of global risk appetite is zero.

o Scale/volatility of amplification: more than 3-fold exaggeration of the underlying fundamentals, end-2011.

16

THE ECONOMIC FUNDAMENTALS

• Competitiveness measured by unit labour costs

• Gov’t debt/GDP, linear and non-linear, so very high debt/GDP has larger effects on spreads

• Private debt/GDP

• Fall-out from housing market crises, worse for countries with previously high hp growth

• Short-term news on growth and inflation

-all relative to Germany.

17

Comparing actual & fundamental spreads for Italy

Composition of fundamental spread for Italy

18

Comparing actual & fundamental spreads for Spain

Composition of fundamental spread for Spain

19

How do these fundamentals affect future growth?

o Remarkable findings from a 9-country panel model (with fixed effects) for 2 years ahead growth rate of real GDP relative to Germany:

o Competitiveness, 2-year lag

o Gov’t debt/GDP has +’ve growth effect for moderate debt, -’ve for very high debt, 2-year lag

o Private debt/GDP has –’ve effect, 1-year lag

o Current year hp growth has +’ve effect but overwhelmed by –’ve effects of appreciation in previous 2 years

20

…..future growth

o Inflation in previous 3 years has negative effect, reinforcing story about competitiveness

o Tendency for partial reversal of high growth rates in previous 3 years.

o All highly statistically significant effects.

21

Policy implications

o Competitiveness and low relative inflation really matter :– emphasis on labour and product market reforms, productivity, wage restraint.

o Powerful housing market overshooting effects and negative effects of high private debt/GDP:- good financial regulation and macro-prudential policy are crucial.

o Fiscal austerity is bad for growth except for extreme levels of govt debt/GDP.

o For a growth promoting strategy, the weights on the economic fundamentals driving yield spreads overemphasise govt/debt, except for extreme levels.

22

Conclusions

o A new European monetary and fiscal authority (EMFA) could set weights on the four fundamental economic factors driving risk spreads on 10-year bond yields combining econometric evidence from models for spreads and models for relative growth rates.

o These ‘shadow prices’ would incentivise governments (and prudential regulators) to promote reform agendas for improved competitiveness and credit and housing market stability, and avoiding excess govt debt – without a central fiscal authority.

23

Change policy, in the absence of eurobonds with risk

spreads…

o Current overemphasis on fiscal austerity – better a general fiscal expansion (but for the highest govt debt/GDP economies)

o ‘QE for the people’ (http://www.voxeu.org/article/combatting-eurozone-deflation-qe-people )or explicit monetary financing of government spending is appropriate in a deflationary world in which the euro-area, with its large trade surpluses, is sucking demand out of the world economy.

o Within the euro-zone, need to reward governments and unions promoting labour and product market reform.

24